Since 1996, the National Environmental Research Institute (NERI), Denmark, has developed a comprehensive and unique integrated air pollution model system, THOR. The model system includes several meteorological and air pollution models capable of operating for different applications and different scales. The system is capable of accurate and high resolution three-days forecasting of weather and air pollution from regional scale over urban background scale and down to individual street canyons in cities - on both sides of the streets. Coupling models over different scales makes it possible to account for contributions from local, near-local as well as remote emission sources in order to describe the air quality at a specific location - e.g. in a street canyon or in a park. The system is used in connection with the urban and background monitoring programs in Denmark. Furthermore, the system can be used to forecast air pollution from accidental releases as e.g. power plants, industrial sites and natural or human made fires.

The main purposes of the THOR system are forecasting, nowcasting, emission reduction scenarios, retrospective analyses and air pollution assessments and management. The system can be used for information and warning of the public in cases of high air pollution levels and for policy management (e.g. by emission reduction or traffic scenarios) of many different chemical compounds. The system can be applied operationally for any location all over the world. The system consists of several different air pollution models - all developed at NERI during the last decades. A schematic diagram of the different modules and the data flow chart of the THOR system is shown in the figure below. The model system consists of a coupling of several models, briefly described in the following.

Present capabilities of the THOR system include all aspects within forecasting, nowcasting, supplement to monitoring programs, scenarios, retrospective analyses, assessment and management of air pollution. Some examples are:

1. Three-day high-resolution regional weather forecasts.
2. Three-day regional air pollution forecasts of 56 chemical compounds, e.g. ozone, sulphur, nitrate, particles, etc.
3. Three-day urban background air quality in specifically identified cities.
4. Three-day urban air quality forecasts at street level - at both sides of the streets.
5. Three-day forecasts of accidental releases into the atmosphere from e.g. nuclear power plants, fires, chemical industries, etc.
6. Emission and traffic reduction scenarios for air pollution management and decision making.
7. Multiple-point and area source dispersion modelling, for determining the effects on air quality caused by proposed new emission sources (e.g., new power plants, chemical industries, commercial activities).
8. Automated production of data, visualizations (maps and time series), information and warnings.
9. Data, forecasts and warnings are disseminated to the authorities and decision makers.
10. Data can be disseminated to the public via Internet or other media.

A three-dimensional numerical weather forecast model, Eta, is applied. This model is initialized with data from a global circulation model, run at the National Centers for Environmental Prediction, NCEP, USA. Data from this global circulation model are the starting point for nearly all weather forecasts in the USA, and for many forecasts in Europe (e.g., Belgium, Greece, Yugoslavia and Iceland). The spatial resolution of the weather forecast model is e.g. 39 km x 39 km over the global grid and 10 km x 10 km over a sub-domain (see the two figures below for an example). Three-dimensional information on winds, temperature, humidity, clouds, precipitation, turbulent fluxes, radiation, etc. can be visualized e.g. every six hours as maps and e.g. every one hour as time series for specific locations. The figures below show the precipitation and surface pressure on November 12^th, 2002 for Europe and Denmark.

The weather forecast is used as input to a long-range transport air pollution model, the Danish Eulerian Hemispheric Model, DEHM, producing air pollution forecasts on regional background scale (e.g. the greater European scale). The operational version of the model calculates transport, dispersion, deposition and chemistry (including photochemistry) of 56 chemical compounds. Furthermore, the model can be used to describe and forecast sand/dust storms. The emission data used in DEHM are derived from a combination of information provided by the European Monitoring and Evaluation Programme (EMEP) and global emission databases. The two figures below show concentrations of nitrogen-dioxide over Europe and Denmark on November 12^th, 2002.

Meteorological data from the weather forecast and air pollution concentrations from the long-range transport model are subsequently used as input to the Urban Background Model, UBM, calculating the urban background air pollution based on emission inventories with a spatial resolution down to one kilometer. The model is e.g. run operationally for the central city of Copenhagen and for the city area of Aalborg, Denmark, and the results of these calculations are published on the Internet four times each day. The UBM model, in the version presently applied in Denmark, is suitable for calculations of urban background concentrations when the dominating source is the road traffic and/or large point sources. The two figures below show concentrations of nitrogen-dioxide over the city of Aalborg (upper figure) and the city of Copenhagen (lower figure).

The output from the urban background model is used as input to the Operational Street Pollution Model, OSPM, producing the air pollution concentrations at street level at both sides of the streets in cities. The model calculates air concentrations of NO, NO₂, NO_x, O₃, CO and benzene in the street canyon at both sides of the street. Particles will be included in the model in the near future. The OSPM has been successfully tested under specific European field campaigns in a variety of different climatic and air quality conditions in, e.g., Copenhagen, Gothenburg, Helsinki, Oslo, Brussels, Berlin, Hanover, and Milano. It has also been tested and applied in Beijing, China, under a cooperation agreement with Tsinghua University.

Due to the circulation of air in street canyons (see the figure above), the air pollution concentrations can be very different at the two sides of a street. This is illustrated in the two sets of figures below. The upper set of figures show a three-day forecast of air pollution concentrations at the eastern and western side of a street in Copenhagen for different chemical compounds. Depending on the meteorological situation, the concentration levels are very different. In the lower set of figure, the maximum value of the two sides of the street is visualized as colored levels. Blue indicates concentrations below mean, green indicates mean concentrations, and red indicates air pollution concentrations above mean.

-->

In addition to urban air quality forecasting, the multi-point (plus area source) dispersion model, OML, has been integrated into the THOR system. This new feature is based on the coupling of the OML point and area dispersion model to the urban background model (UBM). The OML model is the standard model for routine regulatory applications according to the guidelines issued by the Danish Environmental Protection Agency. Thus, it is used for estimating the optimal heights of industrial stacks. OML is a local-scale operational air pollution model for estimating dispersion of a passive, or possibly buoyant, gas from strong point and area sources. It can be applied to distances up to approximately 30 km from the source. In the figure below, the air pollution from two distinct sources in the city of Aalborg is shown as an example.

Furthermore, the weather forecast drives the Danish Rimpuff and Eulerian Accidental release Model, DREAM, used in connection with accidental releases at greater scales as e.g. the Chernobyl accident. DREAM is a combined Lagrangian and Eulerian model, where the Lagrangian part handles the initial near-source transport and dispersion (up to ~300 km from the source) and the Eulerian part calculates transport and dispersion in an area covering e.g. Europe. The model can be used for any accidental release from power plants, industrial sites, natural and human made fires, etc. The first of the two figures below shows the radioactive concentrations of ¹³⁷Cs, 10 days after the Chernobyl accident. The second figure shows the total deposition of ¹³⁷Cs some weeks after the Chernobyl accident.

The entire THOR system is currently run operationally, up to four times every day, initiated with data at 00 UTC, 06 UTC, 12 UTC and 18 UTC. The system is fully automated - meaning that the entire procedure of receiving the data, running the models, producing the visualizations and sending the specified results to the end-users is controlled by automated procedures. The operational performance of the system is monitored every day. The whole system and the operational procedure have been run, tested and validated since August 1998.

Example of comparison of some model results with measurements as time series for January 2000 is given in the two figures above. The upper figure shows a comparison of hourly values of measured meteorological parameters in the city of Aalborg, Denmark, and the weather forecast. The lower figure shows a comparison of hourly values of measured concentrations and the air pollution forecast using the Operational Street Pollution Model at the street Vesterbro, Aalborg. As seen in the figures, the models produce very accurate forecasts of both weather and air pollution parameters.

All weather and air pollution data from the system can be disseminated to the authorities, decision-makers and the public. The raw data can be displayed as maps or as time series. Furthermore, information about exceedances of critical air pollution levels can be extracted and displayed as color codes or given as compressed information, as e.g. "below mean", "mean", "above mean", "high" or "warning". An example is given below for the inner city of Aalborg, Denmark. The home page (in Danish) displays the air pollution at ten streets in the city center, which are colored according to the five possible information levels described above. Furthermore, time series including the air pollution concentrations are given below the city map for detailed information.

Jørgen Brandt, PhD, Group Leader, Senior Scientist, Geophysicist
National Environmental Research Institute, Department of Atmospheric Environment
Frederiksborgvej 399, P.O.Box 358, DK-4000 Roskilde, Denmark
E-mail: jbr@dmu.dk
Phone: +45 4630 1157
Fax: +45 4630 1214

For a demonstration of some of the output from the THOR system: http://www.dmu.dk/AtmosphericEnvironment/thor/index.htm

Link to the system operated by the city of Aalborg (note: If you get an error message, then click on the button "Luftforurening").

The National Environmental Research Institute (NERI) is a sector research institute under the umbrella of the Danish Ministry of Environment. NERI's mission is to provide the scientific basis and environmental data necessary to reach decisions on environmental matters where political, administrative, and commercial activities each play a role.

NERI performs strategic and applied research and is responsible for collection and the overall management and co-ordination of data on nature and the environment. As a research institute, NERI is separated from the political/administrative system when carrying out research and monitoring. As one of NERI's ten departments, the Department of Atmospheric Environment (ATMI) has around 70 employees and carries out broad-based research on air pollution, air chemistry, deposition to both land and water, modelling, forecasting, and system analysis.

ATMI participates in several national, EU, and other internationally funded projects, in various aspects of urban air pollution measurements, atmospheric modelling of meteorological and air pollution phenomena, and forecasting; pesticides and mercury; and air-surface exchange of nutrient compounds. NERI's atmospheric activities involve a long tradition for monitoring with both measurements and models, and for research dedicated to the improved performance of monitoring systems for use in both Denmark, Greenland, and abroad. As part of NERI's international obligations, NERI has implemented protocols for data collection and quality control for the Danish air quality monitoring programme as well as holding the national database of the monitoring program. Finally, NERI is responsible for evaluating the state of the atmospheric environment, assessment of recent trends, and future projections.

NERI's Department of Atmospheric Environment has substantial experience in the development of atmospheric assessment, forecast and warning systems, for use all over the world. These activities have been carried out in collaboration with a variety of international organisations. Since 1998, the department developed an atmospheric dispersion modelling and forecast system (including air pollution). In previous projects, three-dimensional simulations for use on text TV and television were exploited as a next generation display system for government authorities and the public. As a follow-on, an integrated weather and air pollution system was developed during late 1998 and steadily improved through 1999, which spanned global down to urban street scales. This system, THOR, utilises data off the Internet to drive a series of nested systems; as a result, it is a very low cost approach in producing services to a variety of environmental sectors. The system and the approach behind it are fully portable. A validation procedure has been in place since late 1998, with excellent performance. The system was accepted into the background monitoring program of the Danish Environment during 1999 (in support of EU requirements), and it is presently used to support a variety of projects across Europe and the sub-Arctic.

Web sites: www.dmu.dk or www.air.dmu.dk